HP 5316A Teardown and Frequency Standard

[David Hess]

Is there any reason why the 5V OCXO can not be powered directly by the existing +5.00V supply (pin 4)?

To me it seems that the +5.00V (pin 4) is powered by the unregulated +7,5V that is also present on pins 1 and 6.

The Morion MV85 OCXO has a peak power consumption of <600 mA during warm-up.

The way it was originally designed, the XO and TCXO use the +5 volt logic supply but the OCXO instead has a discrete low dropout regulator to provide +5.0 volts from the +7.5 volt unregulated supply.  Offhand I can think of two reasons HP did it this way:

• This distributes the power dissipation across two regulators and avoids overloading the discrete +5 volt logic regulator.  Doing it this way would be less expensive if a majority of the counters did not include the OCXO option because the +5 volt logic regulator would not need to be as powerful.
• Noise and drift on the logic supply may compromise the OCXO performance.

I probably would have used a separate regulator even for the XO and TCXO.

[nfmax]

The OCXO is powered from the permanently on 7.5V supply, so it is kept running in standby. Crystal oscillator drift is approximately proportional to the square root of time, so the longer it has been powered on the (relatively) more stable it gets.

[David Hess]

That is an even better reason that I should have remembered.  My best counter does the same thing even with the TCXO.

[amc184]

Yeah, you guys have got it.  The two main reasons are:
- To keep the OCXO powered and up to temperature at all times so the counter can be turned on and used without delay.  An SC cut crystal won't even put out the right frequency when it's cold, let alone stability.
- To make sure the 5V rail isn't overloaded.  An OCXO draws a lot of current.  The MV85 is actually quite frugal, drawing up to 600mA on start up and 200mA steady state, many are much worse.  You can't assume a regulated rail can provide this much extra current.

Besides all of that though, when I designed the board I just wanted to duplicate the function of the original as closely as possible with more modern parts that are easily available now.  I didn't put much time and thought into changing anything, this was just a quick side project.  I mainly do audio.

[HiTech]

Very nice work/detail on your mod -- AMC184.   

[JoeD]

Received a board from amc184 and have built and fitted to my 5316a. Working well, have used rubidium reference to set it up and it's very stable. The Morion OXCOs are absolutely rock steady and the fourth picture shows the oscillator output reading on a Racal 1998 in overflow mode. That's 10.000000010 MHz ie 10 parts in a billion out...
Nice board Anton, thanks very much

JoeD

[nixxon]

Received a board from amc184 and have built and fitted to my 5316a. Working well, have used rubidium reference to set it up and it's very stable. The Morion OXCOs are absolutely rock steady and the fourth picture shows the oscillator output reading on a Racal 1998 in overflow mode. That's 10.000000010 MHz ie 10 parts in a billion out...
Nice board Anton, thanks very much

JoeD

Very nice results and heat-sink. Did you do the math to calculate the minimum size of the heatsink for the LM1085? With a Vin to the LM1085 of ~9V and a 5V output, the voltage difference is 4V. Multiply that with the < 600 mA power consumption of the OCXO, and it seems that the LM1085's output section must dissipate some 2.4W during OCXO warm up. After a couple of minutes the power consumption of the OCXO is <200mA, so the LM1085's output section needs to dissipate up to 800mW. In addition to the output dissipation, the LM1085 also has to dissipate the power from the control section. This is possibly negligable compared to the power dissipation?

The next step would be to calculate the thermal resistance between Junction and Ambient. I guess any value < 30 deg C/W would be safe...

[nixxon]

I forgot to order a heatsink for the LM1085 voltage regulator when I ordered all the parts for the OCXO assembly. I found a disposable serving tray that was made by thin (~0.3 - 0.4mm) Aluminium and used scissors to cut out two identical shapes that could be folded into a suitable heatsink.

Then I could finally finish the last part of OCXO assembly by installing the LM1085. Little did I know that the OCXO would fail miserably... 

[nixxon]

Then I could finally finish the last part of OCXO assembly by installing the LM1085. Little did I know that the OCXO would fail miserably... 

When I slowly increased the voltage to the 004 equivalent assembly, I noticed that the power supply seemed to be faulty. The voltage didn't go up as much as expected. But the amps did, rising to ~3A. And I could tell that the heatsink I had made was getting really warm. So I measured the resistance between the assembly voltage input pin and GND, and it was above 100 Ohms and not very low/shorted. However capacitor C105 was shorted. And so were the other capacitors between the 5V rail (LM1085 OUT) and GND, C104, C106 and C107. Then I looked at the LM1085 datasheet and noticed that the mounting tab, where the 3mm mounting bolt goes through, was not GND, but voltage output. Therefore I removed the bolt going through the LM1085, the heatsink and the PCB. I was not convinced this would make a difference, as the PCB ground plane was not exposed around the hole for the LM1085 on either side. And everyone else that have built his thing have just put a metal bolt straight through. Nevertheless, the OCXO assembly was now reacting very differently as I applied power to it. I could see the 10 MHz output on the scope as well, as I noticed that the current, that this time had started out at ~0.6A, was decreasing. Success!

With the assembly running, I reinserted the 3mm bolt, and all was OK initially.  I had to tightened the nut a little bit before the output of the OCXO was affected. Maybe the designer, amc184, or someone else that can read the gerber design files in reply #12 in this thread can figure out if the bolt hole is drilled through the copper ground plane of the PCB. Maybe it barely makes contact with the ground plane edge in the hole as the nut is tightened?

Now I only have to find a way to fasten the LM1085 to the rest of the assembly.  It is odd that everyone else that have done this job before me seem to have had success by simply putting a plain bolt straight through, without electrically insulating it.

By the way: the 20 turn trimpot is a very good design choice (made by amc184) to adjust the frequency output of the OCXO. No need to have the right tongue angle, as one full turn makes less than 0.1 ppm (1Hz) change.

[nixxon]

5316A option 004 equivalent install (in 5315A) update:
To avoid the problem with shorting the TO-220 style voltage regulator to ground, I replaced the metal screw with a nylon screw.  I ordered a bag of pretty nicely finished M3 nylon screws from dealextreme.com. I orded rather long screws and trim length as needed with side cutters.

At the same time, I ordered a bag of M3 nylon standoffs for the 004 equivalent board. The nylon threads will fasten ok and not damage the 6-32 nuts which are riveted to the counter's main board. The standoffs were too long, so I had to trim the lenght.

Just like amc184 suggested previously, I had to modify the 004 equivalent board (I used a Dremel sanding drum) to accommodate the large plastic pillar in the 5315A. Because of the mod, I also had to  route the RF output trace with a (bodge) wire, as you can see on the picture.

I calibrated the OCXO just before Christmas, and it is still spot on, as seen in the picture. In the picture, the most significant digit ("1") is missing due to overflow mode. The display should be read as 1.00000000 Hz. (I used a Rubidium Standard FE-5680B with 1 pps output and long gate time to calibrate and check calibration).

Thanks to amc183 for making this mod possible.

I also checked how accurate my Siglent SDG1020 signal generator is with a Vanguard 0.1 ppm TCXO installed. It is exactly 0.1 ppm off. A 10 MHz output signal is measured to 9.9999990 MHz on the 5315A. (This is 180 times more accurate than the standard SDG1020 crystal that was 18 ppm off).

[nixxon]

Received a board from amc184 and have built and fitted to my 5316a. Working well, have used rubidium reference to set it up and it's very stable. The Morion OXCOs are absolutely rock steady and the fourth picture shows the oscillator output reading on a Racal 1998 in overflow mode. That's 10.000000010 MHz ie 10 parts in a billion out...
Nice board Anton, thanks very much

JoeD

If I have got it right, the Racal reading of 10.000000010 MHz is only 1 part in a billion out (0.01 Hz or 10 mHz off).
A 3.6° (1/100) turn on the 20-turn pot, roughly,  should make it 10 times more accurate. That is unless the OCXO drifts a lot at this level of accuracy resolution.

[schopi68]

Ugh.  I've just finished my OCXO-Replacement for the 5316B, built some spare devices and now i found this thread. Hmmm... maybe i should have had a look into this forum during the last months.

[nixxon]

Ugh.  I've just finished my OCXO-Replacement for the 5316B, built some spare devices and now i found this thread. Hmmm... maybe i should have had a look into this forum during the last months.

Did you design your own OCXO "option 004 equivalent"? How did you solve it? Any pictures?

[schopi68]

Did you design your own OCXO "option 004 equivalent"? How did you solve it? Any pictures?

yes, i did my own design, starting last november. Currently i am in the progress of writing down the most important things on my web page to document everything as i have the intention to make the device also available for other interested hams and electronic enthusiasts.
In the meantime the first four devices will get a "one month burn in" to ensure that the old MV85 OCXOs i used (in three devices) are still reliable and to give the new AXTAL OCXO in the fourth device some pre-aging before doing the final calibration with my GPS-based reference.
I have some pictures on my website
(This is the overall project page which i am still updating).

This is an image of my device side by side to the original hp 05316-60008 oscillator:

[xrunner]

No need for a new thread - just got this in from an Ebay seller. Got a replacement knob from another seller too. Haven't turned it on yet.

[schopi68]

No need for a new thread - just got this in from an Ebay seller. Got a replacement knob from another seller too. Haven't turned it on yet.

seems to be in a good condition. What is the serial number (to get the production date)?

[xrunner]

No need for a new thread - just got this in from an Ebay seller. Got a replacement knob from another seller too. Haven't turned it on yet.

seems to be in a good condition. What is the serial number (to get the production date)?

2204A09671

Can you get the date?

[PA0PBZ]

Let me throw in a guess: Week 4 1982, made in USA?

[schopi68]

2204A09671
Can you get the date?

yes, that's possible. But i have to correct myself a little bit: the date encoded in the serial number is the date of the last design change. In most of the cases this corresponds to the real production date within a rage of 1 or 2 years.

The serial number is built of two parts:

The serial number prefix contains the date of the design change and the country of origin.
The first two digits indicate the year (beginning with 1960 = 00)
The second two digits indicate the week.
The Letter designates to A=USA G=Germany J=Japan U=United Kingdom
The third part is the 5-digit serial number starting with 00101

So your unit was built after week 4 in 1982 in America.

To get a more exact date you should have a look on the datecode of the parts in the device.

[xrunner]

The serial number is built of two parts:

The serial number prefix contains the date of the design change and the country of origin.
The first two digits indicate the year (beginning with 1960 = 00)
The second two digits indicate the week.
The Letter designates to A=USA G=Germany J=Japan U=United Kingdom
The third part is the 5-digit serial number starting with 00101

So your unit was built after week 4 in 1982 in America.

To get a more exact date you should have a look on the datecode of the parts in the device.

Thanks for that info.

Since I'm doing a cosmetic restore, I'll post a few more pics that I didn't see in the thread. I have turned it on and it's working properly. I did find the microcomputer IC was not seated properly in it's socket, so fixed that. Also interesting is the LED display ICs are all in sockets, as opposed to the two other HP DMMs of that era I have, which are all soldered in. Noted too is the ability to slide the plexiglass panel out from the aluminum front panel. Also found was the B Channel Level / Sens pot had a leg worked loose from the solder, so needs to be touched up.

[amc184]

yes, i did my own design, starting last november.

Your design looks great schopi68.  It's funny how close the two independent designs came out.  I chose to make mine horizontal because I thought it would be hard to get the right angle brakets that the original used.  I like the unusual standoffs you've used, are they an off the shelf part?

So your unit was built after week 4 in 1982 in America.

That's not quite how the HP serial numbers work.  The first two digital are years since 1960, but it's the year of the last design revision.  That's why that units serial starts with 22, meaning 1982, but many of the ICs are date coded 1983.  It was a version they started producing in 1982, but this unit was from 1983 or later.

[schopi68]

Your design looks great schopi68.  It's funny how close the two independent designs came out.  I chose to make mine horizontal because I thought it would be hard to get the right angle brakets that the original used.  I like the unusual standoffs you've used, are they an off the shelf part?

Yes, tt is quite interesting to see how many OCXO-Boards for these old HP counters are currently on their way. Maybe now the time is ripe for this kind of boards. At the beginning of the year i planned to design a board for my HP 5334B too after finishing the 5316-pcb. But i am no longer sure if somebody else will do this job faster.

The standoffs are off the shelf parts, which i had to modify slightly. Normally they are mounted in a 90° direction, with the flattened side to the pcb. There has been a threaded hole in this side which i had to drill out. With this modification (and somewhat longer screws) it fits perfectly to the counters main board.

So your unit was built after week 4 in 1982 in America.

That's not quite how the HP serial numbers work.  ...

Pleas have a second look to the description i made - i think i already wrote this.   

[schopi68]

Since I'm doing a cosmetic restore, I'll post a few more pics that I didn't see in the thread. I have turned it on and it's working properly.

Great to see that it's still working!

When the device is still open (and maybe reassembled again  )... I'd like to know if my pcb could fit into this device or if i have to find another heatsink. Could you measure the distance from the oscillator pcb's surface to the case?

[xrunner]

When the device is still open (and maybe reassembled again  )... I'd like to know if my pcb could fit into this device or if i have to find another heatsink. Could you measure the distance from the oscillator pcb's surface to the case?

Yes I will do that today.

Also have a question about the PS filter caps, I'll post a pic about that as well.

[xrunner]

The distance from main board to the top is 33 mm. Distance to the side of case is 28 mm (the case is curved so this is the smallest distance.

Have a question also. There are two main filter caps. Both of them have one negative lead and two positive leads. The 4700 uF axial has the two positive leads going to two two different traces, and the 2200 uF radial has the two positive leads going to a single trace (but on the schematic it does not indicate this). If I might want to replace these going forward, what is the designation to specify in the parts search engine of places like Mouser? I am not familiar with caps like these with two positive leads that aren't identified on the case as such. Does that means their are two complete capacitors inside them even though it does not specifically say so on the outside?